finance function cost as percentage of revenue

What a Finance Function Costs: Benchmark Statistics (2026)

What a Finance Function Costs: Benchmark Statistics

Last updated: September 2026

The benchmarks that answer "what should our finance function cost" exist, but they are scattered across paywalled programmes and a handful of signed articles published years apart. This page collects the figures that are publicly traceable, with the source, the year, and the sample size for each. Where a sample size was not published, it says so.

Finance cost as a share of revenue

1.0% of revenue is the median total cost of the finance function. Top performers run at 0.66% of revenue and bottom performers at 1.5%. From APQC cross-industry benchmarking, reported January 2025. The article does not state the sample size for this edition, which places it below the other APQC figures on this page in evidential weight. (APQC, via Perry D. Wiggins, CFO.com, January 2025)

1.2% median, 0.7% top quartile, 1.8% bottom quartile. An earlier and better-documented cut of the same APQC programme, based on 1,515 organizations, published January 2018. (APQC Open Standards Benchmarking, via Marisa Brown, CFO.com, January 2018)

0.4% vs 0.9%: scale changes what "good" means. Among top performers, organizations above USD 1bn in revenue ran finance at 0.4% of revenue, against 0.9% for those below USD 1bn. Comparing a small company against an unsegmented benchmark will therefore understate its performance. Same APQC cut, n = 1,515, January 2018. (APQC, via Marisa Brown, CFO.com)

0.55% of revenue among top quartile organizations. PwC states in its 2024 Finance Effectiveness Benchmarking Study that "Top Quartile organizations have reduced the cost of finance to 0.55% of company revenue". The study draws on nearly 1,000 finance benchmarks, updated with roughly 200 studies run over the 2021-2023 benchmarking period, on companies with USD 1bn or more in revenue in industrialised countries. This figure is stated in the report text; other values in that report appear only in charts and are not reproduced here. (PwC, Becoming the Catalyst, 2024 Finance Effectiveness Benchmarking Study)

How the spread has moved

0.6% to 2.0%, then 0.66% to 1.5%: the gap has narrowed from both ends. APQC's January 2025 reporting notes that a decade earlier top performers sat at 0.6% and bottom performers at 2.0%. The top has drifted slightly upward while the bottom has improved markedly, compressing the range. Note that these are two separate cuts of a benchmarking programme rather than a tracked panel, so this is a shift in the population measured, not a trajectory followed by fixed companies. (APQC, via Perry D. Wiggins, CFO.com, January 2025)

How many people finance runs on

78.6 finance FTEs per USD 1bn of revenue is the median, on the most recent cut. The 25th percentile runs on 45.5 or fewer, the 75th percentile on 102.1, so the bottom quartile needs more than twice the headcount for the same work. APQC Finance Organization Open Standards Benchmarking, published December 2024. As with the January 2025 cost figure, this edition does not state a sample size. (APQC, via Perry D. Wiggins, CFO.com, December 2024)

69.4 median on the earlier, better-documented cut. Top quartile at 36 or fewer, bottom quartile at 141.6 or more. Same APQC programme, 1,784 entities, published January 2019. Both readings are given because the 2019 edition publishes its sample and the 2024 one does not. Read the pair as an order of magnitude, not as a trend: these are two cuts of a benchmarking programme, not a tracked panel, so the apparent rise in the median reflects a change in the population measured as much as anything else. (APQC, via Perry D. Wiggins, CFO.com, January 2019)

Where the time goes

46% of FP&A time is spent on data collection and validation. FP&A Trends Survey 2025, n = 459 finance professionals worldwide (38% Europe, 38% North America, 8% Asia). The report's own series puts this at 53% in 2019 and "little changed from 2024", while its key-findings summary calls it "the highest figure in five years"; the two do not agree, and we report the series. See the source note below. (FP&A Trends Survey 2025)

31% of FP&A time goes to generating insight and driving action, down from 35% in 2024. Same 2025 edition, which states the drop explicitly. Time allocation is covered in full, with its barriers, on Where FP&A Time Actually Goes. (FP&A Trends Survey 2025)

15% of organizations use fully driver-based models, and 2% dynamic ones where AI adjusts the drivers in real time. A further 40% run partially driver-based models built on static, manually selected drivers. FP&A Trends Survey 2025, n = 459. Not comparable with the 2024 edition's "9% fully automated / 54% partially calculated": the categories were redefined between the two, so the apparent jump is a change of definition, not of practice. (FP&A Trends Survey 2025)

30% have not upgraded their planning tools in five years or more. FP&A Trends Survey 2025, n = 459. (FP&A Trends Survey 2025)

A note on these sources

APQC runs the most widely referenced benchmarking programme for finance cost, but its own measure pages are not publicly accessible. The figures above come from signed, dated "Metric of the Month" articles in CFO.com that name APQC as the data source. Where APQC published a sample size, it is stated; where the article did not, that is flagged inline rather than omitted.

FP&A Trends is a self-selected panel recruited through a professional community and sponsored by a planning software vendor. It publishes its methodology and sample composition, which is why it is cited, but it is not a probability sample and likely over-represents organizations already investing in FP&A transformation. Read those figures as directional.

The APQC cost and FTE figures date from 2018, 2019, December 2024 and January 2025. Finance benchmarking moves slowly, but any figure here older than five years should be treated as an order of magnitude rather than a current reading. Note the pattern across editions: the recent CFO.com articles report the percentiles without the sample size that the 2018 and 2019 ones published. Where both exist, this page gives both rather than replacing the documented figure with the newer undocumented one.

Sources

  • APQC, cross-industry finance cost benchmarking (via Perry D. Wiggins, CFO.com, January 2025). Link
  • APQC Open Standards Benchmarking, finance function cost (via Marisa Brown, CFO.com, January 2018). Link
  • APQC Finance Organization Open Standards Benchmarking, FTEs per USD 1bn (via Perry D. Wiggins, CFO.com, December 2024). Link
  • APQC Finance Organization Open Standards Benchmarking, FTEs per USD 1bn (via Perry D. Wiggins, CFO.com, January 2019). Link
  • PwC, Becoming the Catalyst: How finance functions are driving shareholder value, insights from PwC's 2024 Finance Effectiveness Benchmarking Study (nearly 1,000 finance benchmarks; ~200 studies over the 2021-2023 period; companies at USD 1bn+ revenue). Link
  • FP&A Trends Survey 2025. Link
  • FP&A Trends Survey 2024, cited only as the earlier baseline the 2025 edition supersedes. Link

Excluded figures

  • Hackett Group "Digital World Class" finance figures (45% lower cost, up to 42% fewer FTEs, 68% more time on forward-looking analysis, 99% of journal entries automated, June 2025; and the 2023 series on finance operations costs). No sample size or methodology published, and the values are relative gaps without absolute anchors. The underlying research is gated. Excluded on the context requirement, not on suspicion of inaccuracy.
  • Hackett "AI World Class" projections (order-to-cash costs down 52-59%, FTEs per USD 1bn down 56-64%, July 2026). These are modelled projections rather than observed benchmarks.
  • PwC chart-only values (FTEs per USD 1bn, share of FP&A time in analysis, automation rates). Present in the report as charts without labelled values; extracting them would risk misstating the series.
  • "Finance teams spend 30% (or 50%) of their time gathering data," attributed to PwC. Traced through a chain of blogs citing blogs. PwC measures a different construct, and the quoted figure is not verifiable in the report.
  • "26% of finance departments rely exclusively on spreadsheets." Circulates via software vendor blogs with no identifiable primary source.
  • Gartner figures on finance cost and Excel dependence, including "75% of mid-market finance teams run their primary planning models in Excel" and "by 2026 more than 70% of finance organizations will have abandoned spreadsheets as their primary planning tool." Neither is traceable to a Gartner publication, neither cites a report number, and the second is contradicted by measured survey data. Treated as unsubstantiated.

Changelog

  • 2026-09: Link rot and freshness pass. The PwC source 404'd; re-sourced to the same document (Becoming the Catalyst, 2024 Finance Effectiveness Benchmarking Study) at a live PwC URL, and the methodology line corrected to what the report actually states (nearly 1,000 benchmarks, ~200 studies over 2021-2023, USD 1bn+ revenue) in place of an unsupported "data collected December 2023". Added APQC's December 2024 FTE cut (median 78.6, 25th percentile 45.5, 75th percentile 102.1) alongside the 2019 one, which is kept because it publishes its sample size. Moved the FP&A time figures from the 2024 to the 2025 edition, which reverses the direction of the insight figure (35% to 31%) and redefines the driver-based categories; the page previously carried the 2024 numbers while Where FP&A Time Actually Goes is the page that covers them, so the overlap here is now reduced to a pointer.
  • 2026-07: Initial version. Ten figures across finance cost as a share of revenue, cost spread over time, FTE benchmarks, and time allocation. Hackett Group figures reviewed and excluded for undisclosed methodology; two Gartner-attributed figures reviewed and found untraceable.

Further reading

Compiled by Layerz.

Related reading